Computationally efficient algorithm for photoplethysmography-based atrial fibrillation detection using smartphones.

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Atrial fibrillation (AF) is a common arrhythmia and a major cause of stroke and heart failure.
• Current AF diagnosis relies on electrocardiogram (ECG)-based monitoring, which is resource-intensive.
• Photoplethysmogram (PPG) offers a convenient, self-monitoring alternative for heart rhythm assessment.

Purpose of the Study:

• To develop a low-computational, low-memory PPG-based algorithm for AF detection using smartphones.
• To explore novel statistical features and classification methods for improved AF identification.
• To evaluate the algorithm's performance on clinical data.

Main Methods:

• Modified PPG signal acquisition protocol.
• Exploration of new statistical discriminating features.
• Application of sequential forward selection (SFS) and support vector machines (SVM) for classification.
• Evaluation using receiver operating characteristic (ROC) curves and statistical measures.

Main Results:

• A PPG-based AF detection algorithm with low computational and memory requirements was developed.
• The combination of Shannon entropy and median peak rise height demonstrated perfect AF detection accuracy.
• The algorithm effectively utilizes smartphone technology for arrhythmia monitoring.

Conclusions:

• PPG signals hold significant potential for reliable, early detection of atrial fibrillation.
• Smartphone-based PPG analysis offers a convenient and accessible approach to AF screening.
• The proposed algorithm provides a promising tool for remote patient monitoring and early intervention.